Gemini API adds background execution for agents, direct MCP server access

Google has enhanced its Gemini API with an update to its Managed Agents feature, enabling asynchronous execution for long-running agent tasks. This new Background Execution capability allows developers to initiate tasks without maintaining an open HTTP connection, receiving an interaction ID to poll for status or subscribe to events. Additionally, Managed Agents can now directly connect to remote MCP servers for accessing private databases and internal APIs from within a secure sandbox environment. These updates aim to address common failure points like connection timeouts and network instability, improving the reliability and flexibility of AI agent development. AI

IMPACT Improves reliability for AI agent development by enabling asynchronous task execution and secure access to internal resources.

RANK_REASON This is an update to an existing product feature (Managed Agents) rather than a new frontier model release or significant industry-wide event.

Read on dev.to — LLM tag →

• Background Execution
• Claude
• Gemini API
• generative pre-trained transformer
• Google
• Managed Agents
• MCP
• provod.ai

• product
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →